• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

What does 10000RPM do for performance?

H

Hate_Bot

2[H]4U
Joined
Jan 23, 2005
Messages
3,334
I was wondering, everyone always speaks of 10000rpm lke its a godsend, I was wondering what exactly this did for performance/? Does it help much or at all? also, why are the 10000 drives so small? the biggest Ive seen is like 75 gigs.

whats the point?
 
10000rpm have very fast access times. 3ms compared to 8ms on faster 7200rpm drives. scsi has bigger 10000rpm drives, 140 and 320gb. i think 10k rpm drives are overrated and for the price of a 74gb 10k rpm drive, you could buy a 160+gb 7200rpm drive.

although, if i had mad amounts of money, id get 2x74gb in raid 0 for my OS/programs drive, and 2x400gb drives in raid 1 for personal files.
 
You can achieve nearly comparable performance at 7200RPM, but only with ridiculously tight sector density. The Seagate 7200.8 drives, especially the 400GB ones, have obscene access times, even more so when you use an NCQ capable controller.
 
Depends on what 10K drive. The 36GB Raptor goes down in defeat to a 16MB Maxtor drive, and is merely equal to a 250GB Hitachi w/ 8MB buffer. On the flip side, the WD740GB with some firmware tweaks and a slightly more powerful actuator compared to the 36GB, stands head and shoulders above any 7200 RPM. Same for SCSI drives - a 10K SCSI drive from three generations ago is significantly slower than most any current 7200RPM, while Maxtor Atlas 10K V and Fujitsu MAT demolish even WD740GD, and comparing 7200RPM drives to them is laughable.
 
Whats the point for any upgrade then? You got clueless people paying an extra $500 for SLI setup to get one higher resolution and AA setting that is only necessary on a few games. People spending $100 more for a faster DVD writer just to burn a DVD 5 minutes faster. people spending $800 more for the newest processor to encode video a few minutes faster, ok you get my drift. Yes you will only see about 20% real world improvement from a 10k Raptor or Cheetah, but considering that the HD is one of the backbones of a computer its a pretty sensible upgrade. And using just one for OS and games, and a larger 7200 for media, backup, and general storage makes pretty darn good sense to me.
 
To put it simply, the faster the platter spins, the faster the head can reach the requested data. The faster it reaches the data the faster it fills the buffer. A full buffer can saturate an interface at speeds in excess of 150MB/s depending on the drive. So, needless to say the faster you can fill the buffer, the more the drive can saturate the interface and thus the more and faster the data gets dumped onto the bus to be moved across the bridges and into RAM to await processing.

It would be nice to see a SATA 150 drive with 16MB buffer @ 10,000+ RPM. That would be killer.
 
Increasing the rotational speed improves both seek and read/write performance. A RAID 0 or 5 setup with 7200 RPM drives might match or beat the read/write performance of a 10,000 RPM drive, but no matter how many drives are involved in the array, it will never have comparable seek performance to a 10,000 RPM drive. Fast seeks improve overall system responsiveness, and is especially evident in audio/video editing scenarios.
 
Suggesting dual raptor 740' s in raid -0 is not allowed in any forum any longer so cut it out... It is also phenomonally stupid... much like most people who dont understand the point of raid.
 
ferrisnox said:
Suggesting dual raptor 740' s in raid -0 is not allowed in any forum any longer so cut it out... It is also phenomonally stupid... much like most people who dont understand the point of raid.
Click to expand...

its amazing how long it takes for incorrect assumptions to be finally weeded out.
 
ferrisnox said:
Suggesting dual raptor 740' s in raid -0 is not allowed in any forum any longer so cut it out... It is also phenomonally stupid... much like most people who dont understand the point of raid.
Click to expand...

So you are saying Raid 0 provides zero performance benefits? :rolleyes:

 
As someone who runs raptors, and knows what raid means....redundant array independant disks. In a server enviornment this is used for one purpose only... fault tolerance... not performance... Hook up as many raptors you want in raid, and I'll beat them every time with a single scsi disk... See what I'm getting at is when you have the money to spend and you want speed you buy the fastest disk technology... otherwise while I admit raid-0 can improve performance in raid-0 (not true raid anyway) in a server enviornment its is more cost effective to buy a single disk that will destroy performance wise any raid... because you want performance and longevity... and raid-0 cuts your longevity probability in half.

On a desktop system... the performance between running a single raptor and two in raid-0 is stupid... this has been tested many times...
 
MeanieMan said:
So you are saying Raid 0 provides zero performance benefits? :rolleyes:

Click to expand...


Eh, he thinks it is stupid because it is RAID 0, not because of the drives in question. Some people, though, like scratch areas with high throughput, you know, people who do a/v editing, or for photoshop batching. His points, I believe, are that the performance gains from RAID 0 are not worth it, and in most cases he would be correct, but not all.

RAID 0 with raptors has its place, its just that he isn't familiar with that place.
 
eastvillager said:
Eh, he thinks it is stupid because it is RAID 0, not because of the drives in question. Some people, though, like scratch areas with high throughput, you know, people who do a/v editing, or for photoshop batching. His points, I believe, are that the performance gains from RAID 0 are not worth it, and in most cases he would be correct, but not all.

RAID 0 with raptors has its place, its just that he isn't familiar with that place.
Click to expand...

Yes. For OS/boot drives, or gaming, all you are really doing is doubling your chance of failure. For multimedia editting of large files, spanned disks have some advantages.
 
ferrisnox said:
As someone who runs raptors, and knows what raid means....redundant array independant disks. In a server enviornment this is used for one purpose only... fault tolerance... not performance... Hook up as many raptors you want in raid, and I'll beat them every time with a single scsi disk... See what I'm getting at is when you have the money to spend and you want speed you buy the fastest disk technology... otherwise while I admit raid-0 can improve performance in raid-0 (not true raid anyway) in a server enviornment its is more cost effective to buy a single disk that will destroy performance wise any raid... because you want performance and longevity... and raid-0 cuts your longevity probability in half.

On a desktop system... the performance between running a single raptor and two in raid-0 is stupid... this has been tested many times...
Click to expand...

You're comparing apples to oranges, a raptor raid vs a 10k or 15k scsi disk.

Take 2 scsi disks, stripe them at a width that is appropriate for your application and it will be faster than the single drive, in any test you care to make with that application.

If you go a little bit farther, what you really do for performance in a server environment, is mirrored stripes. Stripe X number of spindles, then create another stripeset just like it and mirror the two of them. Any decent midrange or topend array will not only do that with no additional overhead, it'll also intelligently split the reads between both mirrors, so you can effectively read from 2X spindles. You of course have to make sure you have a backplane and sufficient controllers to handle the throughput.

The last large server I did was a 44 processor datawarehouse with about 25TB spread across 24 fc ports and 12 midrange arrays. Of course, in that system, most of the activity was reads, so RAID 5 with a nice bit of cache was fine for the data, while result tables/scratch space was put on mirrored stripes.

Is RAID 0 silly on the desktop for 99.9% of the home users? Yes, undoubtedly, but it isn't silly for everything.
 
Gentleman I agree with most of your astute points, I get very upset when people go around screaming that raid- is the solution to faster performance, to everything. My point was about desktop performance.. I just dont think there is any reason that I can think of for raid... , you mentioned video editing that is a valid point but thats not going to be done on a desktop thats gonna be done on a workstation... which is significantly different it all aspects.

But back the the gentlemans question what is the point of 10,000 rpm... the point is speed and on a good disk it is significant over 7200 rpms.
 
eastvillager said:
Is RAID 0 silly on the desktop for 99.9% of the home users? Yes, undoubtedly, but it isn't silly for everything.
Click to expand...
That sums it up pretty nicely.
 
I enjoy my raptors in raid 0 on my computer. I deal with alot of large files, I ripped my entire library of DVDs to to my hard drives, and recording tv shows and stuff too.

But not to get in the arguement about raid...

____

10K rpm is notaceably faster to me over 7200rpm plain and simple. I will never go back to anything slower than 10K now.... which worries me, because we need to see some current gen 10k stuff, like SATA II and NCQ.... who knows... hopefully we will by september when i get off deployment ;) i could care less what comes out until then.
 
05.04.14
think the upcoming solid state flash hard drive, ruged no moving part, no noise, low ele draw and near 0 MS access are the future. Regretably they are only about 4 GB atm, but will certainly increase in size.
Mayby with dedicated OS imbeded memory a seperated 16 or 32MB cache HD using the pcie IO bus will utalize +200MB/s data tfer.
 
"RAID 0 implements a striped disk array, the data is broken down into blocks and each block is written to a separate disk drive
I/O performance is greatly improved by spreading the I/O load across many channels and drives
Best performance is achieved when data is striped across multiple controllers with only one drive per controller
No parity calculation overhead is involved"

RAID 0 is not fault-tolerant and will make loading times(if your a gamer, spawning) much faster then your conventional 7200 rpm drive. People install win xp in under 4 min with a 10,000rpm drive and seek times are 2-3x faster then 7200's. The risk of RAID 0 is that if you lose one drive your data is toast. SCSI is better but this is a cheaper solution.

If I am wrong please correct me.
 
bug said:
05.04.14
think the upcoming solid state flash hard drive, ruged no moving part, no noise, low ele draw and near 0 MS access are the future. Regretably they are only about 4 GB atm, but will certainly increase in size.
Mayby with dedicated OS imbeded memory a seperated 16 or 32MB cache HD using the pcie IO bus will utalize +200MB/s data tfer.
Click to expand...
Flash is pretty slow, unfortunately. Newer usb2 pen drives are only 15 mb/s or so, or they'd be perfect. I fail to understand why manufacturers don't do something like raid0 on these - you aren't as worried about the failure aspect, the speed needs the help, and you certainly aren't worried about price if you're buying flash!

It'd be nice to see nonvolatile ram used for this (eg sram). That would make a good boot disk! :D
 
First link's broken. The others are quite an interesting read.

I didn't see anything about sustained transfer rates in any of those. They're fast, I understand that, especially with no seek latencies, but I didn't think the sequential transfer rates were much more than 20 MB/s. Feel free to make me eat my words ;)
 
killerD said:
Whats the point for any upgrade then? You got clueless people paying an extra $500 for SLI setup to get one higher resolution and AA setting that is only necessary on a few games. People spending $100 more for a faster DVD writer just to burn a DVD 5 minutes faster. people spending $800 more for the newest processor to encode video a few minutes faster, ok you get my drift. Yes you will only see about 20% real world improvement from a 10k Raptor or Cheetah, but considering that the HD is one of the backbones of a computer its a pretty sensible upgrade. And using just one for OS and games, and a larger 7200 for media, backup, and general storage makes pretty darn good sense to me.
Click to expand...

I know I do it because 1) I am 16, I can care less what I spend my money on 2) Two my friends look at my Raid 0 raptors and go "whoa shit, thats awesome" 3) part of computing and tweaking/[H]ardcore rig is having the latest and greatest. Sure it fades but dosen't everything at some point? Treat yourself once in a while, I know I do.
 
unhappy_mage said:
First link's broken. The others are quite an interesting read.
Click to expand...


you must have been pretty fast on that
I generally verify my links
I caught that and replaced it with a wayback machine archive of it both above and in the linkfarm, but you where still too fast for me :p
 
ferrisnox said:
Please just read it.. and think...
http://faq.storagereview.com/SingleDriveVsRaid0
http://www.anandtech.com/storage/showdoc.aspx?i=2101&p=1
If you haven't gotten the hint by now, we'll spell it out for you: there is no place, and no need for a RAID-0 array on a desktop computer. The real world performance increases are negligible at best and the reduction in reliability, thanks to a halving of the mean time between failure, makes RAID-0 far from worth it on the desktop.
Click to expand...


not necessarilly that cut and dried

RAID 0 Hype or Blessing

Femme Taken employs IPEAK SPT traces just like SR does

Because of the striping setup's overhead, the single Raptor WD740GD is faster than the RAID-0 machine, with a queue length of one and a transfer size of up to 32K. When the transfer size increases to 64K however, the two striped Raptors take the lead from the single drive. Increasing the number of operations in the I/O-queue to eight means the striped setup leaves the single Raptor in a cloud of dust.
Click to expand...

on the desktop of the power user p2p-sharing, image- and video editing and backing up all kinds of digital media are frequent activities, too. AnandTech, Storage Review and Tweakers.net all share the same audience: power users and tweakers who want to maximize power, speed and usability of their rig. Unlike the average secretary, our target audience is continuously multi-tasking, which can include downloading everything that can be found on the internet, while taking pictures at the highest possible settings. In short: they demand the highest performance from their desktop.
Click to expand...

Performance scaling in the office and content creation benchmarks performed by Storage Review contrast severely with Tweakers.net's tests, where the Raptor array, as mentioned earlier, showed a 36,1 percent improvement in Business Winstone 2004 and a nice gain of 50 percent in the Multimedia Content Creation Winstone 2004. The fact that Storage Review had the FastTrak S150 TX4 crawling on a 32-bits 33MHz PCI-bus can't be a fair explanation for the disappointing results, since our FastTrak S150 TX2plus was also installed on a legacy PCI-slot. Bandwidth is of greater importance for RAID 0 performance, since the improvements in Storage Review's tests had to be obtained with faster sequential transfer rates. Differences between the Promise FastTrak S150 TX4 and the TX2plus can in no possible way cause such a big difference in results, since both adapters are very similar to each other.

The fact that there is a significant improvement in the Windows XP boot benchmark clearly shows that Storage Review's system could have some obvious advantages with RAID 0 when using correct workloads. Therefore, we suggest that Storage Review's benchmarks are not typical for the demanding use patterns in current office and content creation applications.
Click to expand...

Our observations clearly show that the testing methods of AnandTech and Storage Review are too limited to jump to conclusions about the use of striping in desktop environments - let alone that they could be used to sign RAID 0's death sentence. Both AnandTech's and Storage Review's results of the IPEAK largely contradict Tweakers.net's benchmarks, where RAID 0 does show significant improvements in the I/O performance.

In the case of Storage Review, this can be explained by their use of traces based on older software, which isn't representative of present-day 'power user' workloads. Using system benchmarks to assess storage performance is at the very least questionable anyway, because they mainly depend on CPU performance. The same thing goes for the loading times of games - not to mention the fact that a test limited to two games will never give a realistic view of performance in general.

Most striking is the fact that neither AnandTech nor Storage Review even bothered to test RAID in situations where the performance of the storage system really matters - for example when a backup application, anti-virus tool or compression program is active in the background while demanding content creation applications are running in the foreground. These are conditions that quickly give rise to noticeable latencies and in which single drive setups don't suffice anymore.
Click to expand...

he also has considerably more RAID benchmarking experience than SR ;)

if your going to investigate tweakers.net
breakout world lingo and set it to Dutch ;)

Cliff Notes: it depends on how heavy an access pattern (read multitasking) you employ, like most things its shades of grey not a black and white cut and dried answer, that said you still might see an increase in performance even with heavy multitasking if you have 2 seperate Raptors and a third storage HDD each on thier own channels and your config optimized, your usage pattern is unique, RAID0 isnt a be all answer its also not the shaggy dog its been made out to be recently, deferentiate your true needs for performance space and your storage needs, they arent the same thing ;)
 
Ice Czar said:
you must have been pretty fast on that
I generally verify my links
I caught that and replaced it with a wayback machine archive of it both above and in the linkfarm, but you where still too fast for me :p
Click to expand...
I do read pretty fast. However, the Wayback machine is what gave me the error. Presumably they have problems from time to time.

It makes for interesting reading - if you run a database, I guess. However, I still haven't seen a definitive answer to the "how fast are linear reads/writes" question. Maybe I'm just skimming too much, but I missed it. The 3rd article does say:
RAID 0 performance can be improved using SSDs to stream data from two or more drives in parallel at the n times the bandwidth of the SSD (currently 80MB/sec).
Click to expand...
but that seems to me to be a quick generalization. It could well be accurate, though. Any way you put it, they do appear to be plenty fast. :D
 
Yet another reason to add to the many to disgard Anand to the waste receptacle much of his stuff belongs in. Sounds like SR isn't too far behind him. I'll remember that the next time someone tries to use them in an effort to "disprove" others. :)
 
Ice Czar said:
Cliff Notes: it depends on how heavy an access pattern (read multitasking) you employ, like most things its shades of grey not a black and white cut and dried answer, that said you still might see an increase in performance even with heavy multitasking if you have 2 seperate Raptors and a third storage HDD each on thier own channels and your config optimized, your usage pattern is unique, RAID0 isnt a be all answer its also not the shaggy dog its been made out to be recently, deferentiate your true needs for performance space and your storage needs, they arent the same thing ;)
Click to expand...

I have never (well, never since I've been on the [H]) taken the 'cut and dried' approach on this issue - I've been one of the 'shades of gray' people. I have always maintained that adding spindles, either striped or as separate disks for different tasks, can increase throughput dramatically in some situations. However, I have made it to be the 'shaggy dog' for gamers looking to load their levels faster, and I tell them to buy one Raptor and spend the money that would have gone to a second one on a better video card, better mobo, etc which will definitely boost performance for a gamer, across the board. I also suggest puttign all of their programs/games/OS on the Raptor, and picking up a 7200RPM drive at 1/4 the cost per GB of a Raptor for their 'storage drive' Some people may accuse me of 'hopping on two different bandwagons' or say 'that's confusing, why can't you just have one position and stick with it.' However, as Ice Czar says "your usage pattern is unique" so my answer for a suggested storage setup is unique to every application.

RAID-0 is a lot like SLI - SLI doesn't deliver performance boosts under light loads, and neither does RAID-0. It is only until you dramatically raise the load that SLI begins to show benefits - you don't buy an SLI setup to play in 800x600 - why buy a RAID-0 setup when your disk load will not be that heavy? Also, as another example, why RAID-0 two 7200 RPM drives when one Raptor is probably faster? Most of us [H]'ers understand that a single 6800GT is faster than 6600GT SLI in all but a handful of mostly synthetic cases, why can't we understand that with hard drives? I realize that this is an apples and oranges comparison, but there is a lot of conceptual similarity here, IMO.
 
SirKenin said:
Yet another reason to add to the many to disgard Anand to the waste receptacle much of his stuff belongs in. Sounds like SR isn't too far behind him. I'll remember that the next time someone tries to use them in an effort to "disprove" others. :)
Click to expand...

For workstation work, and heavy desktop productivty work, I can uderstand this truly disk intensive multi tasking model that Ice Czar references in the tweakers.net article. However, this is a gamer website, at least last time I checked. Anyone who is running that kind of disk intensive laod while gaming is an idiot. We're gamers - every FPS, and every eye candy/resolution level counts, and we, at least I hope, are not going to exhibit a usage pattern like Tweakers tests while we're gaming.

Yes, there are workstation users here on the [H] that do the type of work that can benefit from RAID-0, but they are much fewer in number than the gamers. I have no problem recommending RAID-0 to a user who needs a fast scratch disk for content creation, for a user who has a solid backup strategy and needs the highest sequential throughput possible, or a user who beats the drives to death with highly random, high queue work loads that can benefit form having an extra spindle on long stroke seeks. However, none of these scenarios describe a gamer usage model.
 
God bless DougLite (He Gets it)

Well they way I read there wasn't any really questionable content... only shades of gray much like the whole debate... my only point is instead of being so damn polarized "burn all those who raid" and "burn those who do not..." lets come to a consensus... Unless your doing video editing, running a midcost server,or something else its difficult to recommend raid-0 for the desktop.....Other than to just have a giant E-penis
Honestly without a benchmark or an atomic stopwatch do think you could judge the performance difference between a single raptor 740 and two in raid -0... just by using the System????

Honestly you can afford the money to run dual raptor740's in raid-0 then I wouldn't one just get scsi at that point...
admit you have a freakin problem..... I cant help you if you wont admit it
 
DougLite said:
" so my answer for a suggested storage setup is unique to every application.
Click to expand...

exactly and is what start > run > perfmon.msc is for
master that first then start in with your various configs ;)

(addressed to the audience not Douglite of course :p )

I would say that the senerio that Femme Takken points out is a little more broad than the typical DV editing one always used as a caveate, and just basic photoshop with a reasonably large file w\ multitasking would probably qualify is a 2 HDD box

but taking advantage of multiple channels and concurrent access is something that I almost always recommend if you can swing it
 
So your saying you couldnt tell a difference... and I read the tweakers review... and unfortunately while it was good,,, it missed the point entirely....
Too show the advantages of raid-0 they had to process images in photoshop and perform a network backup... Okay here is what I'm saying, that is not desktop usage... so the entire point is moot... we're talking about real life desktop usage...I' have personally never worked in photoshop while performing a network backup... or anykind of backup for that matter.
 
most the gamers I know do a fairly healthy amount of miltitasking in between gaming sessions, with a decent amount of encoding, downloading, surfing and burning happening concurrently

the original point was the assumption that anyones IPEAK SPT traces are representative of yours is false and that the only way to optimize is to regularly monitor and log your accesses and experiment how best to improve them ;)
 
Give me the name of this gamer that plays games, surfs the web, encodes audio and video, downloads files, and burns dvd, and cds.. .concurrently... I would like to meet him or her.. and get some of their magical crack rocks... I'm only asking you to be realistic

I agree raid-0 has a place and a use, just so happens most of the time its not on a desktop computer.
 
you dont get to be a mod without instinctively qualifying your statements
we are as vauge and slippery as the greasist polititian. :p
 
Ice Czar said:
most the gamers I know do a fairly healthy amount of miltitasking in between gaming sessions, with a decent amount of encoding, downloading, surfing and burning happening concurrently

the original point was the assumption that anyones IPEAK SPT traces are representative of yours is false and that the only way to optimize is to regularly monitor and log your accesses and experiment how best to improve them ;)
Click to expand...

In that situation, they would be best suited with a separate P4 system or a dual core system, while gaming on a K8 system. Generally, encoding is CPU limited, everyone knows that downloads are limited by the bandwidth of your Internet connection, and plenty of physical RAM is going to do the most for performance while you're surfing. I will admit that the burning work is definitely disk intensive, and RAID-0 and/or NCQ may make the difference between being able to run a virus scan/open a zip file/etc and burn your DVD at the same time, and may prevent getting a coaster under that heavy of a load. An entirely separate disk for the source of the DVD is going to be a better protection against coasters than RAID-0. Once again, I'm not saying RAID-0 is useless. However, it's applicability on the desktop is rather narrow, and thare are just not many scenarios that benefit form it, particularly when multiple single drives can meet the same goals without the reliability problems of RAID-0, assuming the storage host is up to snuff, and I'd expect any recent Intel or Nvidia chipset to be good to go there.

As for your statement about the benchmarks, it appears that we have reached a conclusion about hard drive benchmarks - they're about as useful as video card benchmarks ;) Useful for comparing drives, but not a good guide on how you should install and configure them.
 
You must log in or register to reply here.
Back
Top